"text filter"

helen

could somebody give me a help regarding text filter. Here is my data loaded in RapidMiner.

key           description
1               loan to help me to start up a new company
2               loan to help to make a new hous
3               loan to help me to pay down credit card debt
4               loan to help me to pay down credit card debt
.                 .................
.                 ................
25             loan to help me to pay down credit card debt
.                 ...................
n               ..................

I'd like to use attibute filter to get the data which match in the description "credit card".  As result sould show me like below:
key           description
3               loan to help me to pay down credit card debt
4               loan to help me to pay down credit card debt
25             loan to help me to pay down credit card debt

should I uese the attribute filter in the preprocessing operater on Root Process? and which category should I chose in the condition_class? should I type "credit card" in the fild of parameter_sting?

many thinks in advance.

land

Hi,
if I understood you correctly, then you want to filter your examples (=rows) and not the attributes (=columns). Hence you would need the example filter.

Here's an example:

<operator name="Root" class="Process" expanded="yes">
    <operator name="ExampleFilter" class="ExampleFilter">
        <parameter key="condition_class"	value="attribute_value_filter"/>
        <parameter key="parameter_string"	value="description=.*credit card.*"/>
    </operator>
</operator>

If you need help on regular expressions, you could take a look in the tutorial or just use google.

Greetings,
  Sebastian

helen

highly appreciated ! Thanls a lot Sebastian  ;D

helen

Could I ask a further question on this? Because sometimes I need to find out in the description some texts, which don’t standing together. E.g if I want to file out "pay debt " in the description. It should give me the same output as above:
key          description
3              loan to help me to pay down credit card debt
4              loan to help me to pay down credit card debt
25            loan to help me to pay down credit card debt
What should I type on the code? value = “description=.*pay. debt*  ???
Thanks a lot in advance

Ryujakk

Hello!

That sounds like you need some more... regular expressions! I learned most of it here: http://www.regular-expressions.info/

For your particular example, I'd use this:

<operator name="ExampleFilter" class="ExampleFilter">
  <parameter key="condition_class"	value="attribute_value_filter"/>
  <parameter key="parameter_string"	value="description = .*pay.*debt.*"/>
</operator>

- R

land

Hi,
I don't know if the mentioned above website contains something equal, but I always found this website helpful:

http://www.fileformat.info/tool/regex.htm

It provides the possibility to test regular expressions on arbitrary text.

By the way: The new RapidMiner 5 contains a builder for regular expressions. For all of us, always forgetting those damned group names...

Greetings,
  Sebastian